Project Summary/Abstract Pancreatic cancer is a leading cause of cancer death. Most patients with this disease are diagnosed at a late, unresectable stage. An effective strategy to reduce the incidence and mortality of pancreatic cancer is the timely identification and optimal treatment of its precursor lesions. One such lesion is intraductal papillary mucinous neoplasm (IPMN) because some of these cyst lesions have potential to progress to invasive cancer. Based solely on clinical and radiologic features, current guidelines for predicting and treating malignant IPMN have a satisfactory sensitivity (>90%) but a dismal specificity (25-30%), compared with final surgical pathology. Such a low specificity has resulted in an unacceptable high false positive rate and hereby a number of unnecessary pancreatic resections associated with surgically-related mortality in benign IPMN. Therefore, there is an urgent unmet need to identify molecular biomarkers to improve malignant IPMN prediction. The long-term goal of the proposed study is to improve clinical IPMN management and prevent pancreatic cancer by identifying risk factors or predictors for malignant IPMN. The primary objective is to evaluate the associations of IR biomarkers and metabolites with malignant IPMN risk. Our central hypothesis that IR biomarkers and metabolite signature can predict malignant IPMN risk has been formulated on the basis of our preliminary data and extensive literature review. We propose to investigate this novel but biologically plausible hypothesis among 400 IPMN patients who have undergone surgery at the Indiana University Pancreatic Cyst and Cancer Early Detection Center. Of these, 118 were classified by final pathology as malignant and 282 as benign IPMNs. Our Specific Aims are: 1. Investigate the associations between selected plasma IR biomarkers and malignant IPMN risk. Selected IR biomarkers are C-peptide, branched-chain amino acids, leptin, high-molecular weight form of adiponectin, retinol binding protein-4, and glycated hemoglobin; 2. Identify plasma metabolites distinguishing malignant from benign IPMN using global metabolomics. More than 800 named metabolites will be measured. Mediation analysis will be run to evaluate whether and to what extent identified metabolites predict malignant IPMN through IR mechanism. 3. Determine the capacity of combining plasma IR biomarkers and metabolites (identified in Aims 1 and 2) with clinicopathologic characteristics for predicting malignant IPMN. The proposed study is expected to identify the IR biomarkers and metabolites that are associated with malignant IPMN risk and demonstrate that the model that integrates IR biomarkers and metabolites with clinicopathologic features is more predictive of malignant IPMN than the model that relies solely on clinicopathologic data. Our expected results are significant because they will inform physicians to make evidence-based clinical IPMN management and open new avenues for preventing this precursor lesion and ensuing pancreatic cancer. Our proposal is innovative because malignant IPMN predictors will be identified from both pathway-based and global-profiling approaches.